The COVID-19 Hospitalization Forecaster uses machine learning to predict the number of hospital beds that will be occupied by COVID-19 inpatients in each state/territory, up to 14 days into the future. To learn more about our methods or access the source code, visit our GitHub repository.
You can also try our API. From the command line, enter e.g.
curl julianrosen.net:5000/?state=AL

The forecasting model was built by Lara Du, Julian Rosen, and Kai Wei as a project for the Erdős Institute's Fall 2021 data science mini-course. Our project mentor was Ben Campbell. The forecast is intended for educational purposes only, we cannot guarantee its accuracy.